1 2 3 4 Previous Next 54 Replies Latest reply: Feb 10, 2010 4:09 AM by g777 Go to original post RSS
      • 30. Re: Is ORACLE slower on Windows then on Linux ?
        108476
        Hi Joel,
        I still prefer "real" unix.
        Me too. You can't beat AIX, HPUX and even Solaris
        I guess I'm in a shrinking minority.
        It's because all the people without jobs can get their hands on is Linux.

        Linux is popular because it's available to the great unwashed, that's the only reason . . .
        I have mixed feelings about this VM stuff.
        Yeah. The whole point of having a mainframe is to have instant sharing of computing resources, why muck it up with all of these resource fencing tools?
        • 31. Re: Is ORACLE slower on Windows then on Linux ?
          108476
          Didn't you ever wonder about Alpha/NT? Used to have one warming my feet in the last century.
          Ha! I had a PDP-11 in my basement once. Got it for free . . .

          It warmed my whole house, but the lights would go dim on my whole block when I turned it on!!!
          • 32. Re: Is ORACLE slower on Windows then on Linux ?
            87944
            It does not take a genius to know that Windows sucks, every Hippie with a Mac knows this . . .
            While Windows do suck, everybody around is eager to hear your comment on your stance:
            Any OS where you can modify the source code dynamically (DLL's) is a VERY POOR design.
            • 33. Re: Is ORACLE slower on Windows then on Linux ?
              108476
              Hi Michael,
              everybody around is eager to hear your comment on your stance:
              I'm not a software engineer, but I have a few in my employ and they that "DLL Hell" is a major reason why Windows sucks!

              I've been told like this:

              http://en.wikipedia.org/wiki/Dynamic-link_library

              "While DLLs are the core of the Windows architecture, they have a number of drawbacks, collectively called "DLL hell"."

              I'll be glad to ask one of my engoneers to pop-in with details if you post the question in their forum:

              http://dbaforums.org/oracle/index.php?showforum=2
              • 34. Re: Is ORACLE slower on Windows then on Linux ?
                Billy~Verreynne
                burleson wrote:

                It's because all the people without jobs can get their hands on is Linux.

                Linux is popular because it's available to the great unwashed, that's the only reason . . .
                Tell us, why Unix kernels (e.g. HP-UX) is a better architecture than a Linux kernel? Or why the memory management and driver model are better?

                Edited by: jkestely on Oct 28, 2009 2:49 PM
                • 35. Re: Is ORACLE slower on Windows then on Linux ?
                  jgarry
                  burleson wrote:
                  Didn't you ever wonder about Alpha/NT? Used to have one warming my feet in the last century.
                  Ha! I had a PDP-11 in my basement once. Got it for free . . .

                  It warmed my whole house, but the lights would go dim on my whole block when I turned it on!!!
                  I still have two. One of them I paid for new (with two 8" floppies and RT-11!, later upgraded to 11/23+), the other rescued on its way to the dumpster. Haven't turned them on in this century, though, kinda afraid what those old bearings will do... I do want to hook one up to the net one day (probably via rs-232 to a pc) and let people know "RSTS is Ready!"
                  • 36. Re: Is ORACLE slower on Windows then on Linux ?
                    jgarry
                    user1148640 wrote:
                    burleson wrote:

                    It's because all the people without jobs can get their hands on is Linux.

                    Linux is popular because it's available to the great unwashed, that's the only reason . . .
                    Well, I may not agree with other things he says or his manner of saying them, but I don't agree that this is bigoted, and I believe to a large extent he is correct.

                    >
                    Try this on for size Burleson.. how about using technical argument in technical discussions?

                    Tell us, why Unix kernels (e.g. HP-UX) is a better architecture than a Linux kernel? Or why the memory management and driver model are better?
                    They are better because they are more uniform within each vendor in their engineering for the specific hardware they run on. Unfortunately, this doesn't entirely apply to the X-86 hardware, but it is unfortunate in general that the X-86 architecture has gained supremacy. Yet another example of marketing and mass appeal winning over technical superiority. Transmeta, anyone? Gee, who worked there?

                    Both the general unix kernels and linux suffer from being around a long time and undergoing evolution and forking, linux less so, but suffering also from the opposite problem of developing too fast. Linus' project management is genius, at least so far, but it does not set the highest priority on developing the best systems for database engines. And there is where Don is exactly right - it's popular because it's available. That doesn't make it good. If it did, Windows would be the best operating system in the world.

                    I'm not even going to address the memory management mess, except to note, in linux it has changed rapidly and still hasn't been gotten right for Oracle databases. Figuring that out is an exercise for the student. Have fun going up and down on the elevator in the swampiness, I mean, swapiness.

                    The driver model sucks completely, you just don't know what is going to work when with what and for how long. Even freakin' winblows, which I'm strongly biased against, is better for that.

                    Edited by: jkestely on Oct 28, 2009 3:35 PM
                    • 37. Re: Is ORACLE slower on Windows then on Linux ?
                      87944
                      jgarry wrote:
                      I do want to hook one up to the net one day (probably via rs-232 to a pc) and let people know "RSTS is Ready!"
                      See SIMH to ride the wave of nostalgia.

                      For example:
                      PDP-11 simulator V3.7-3 [32b data, 32b addresses, no Ethernet]
                      
                      CPU, 11/73, NOCIS, idle disabled, autoconfiguration on, 256KB
                      Throttle = 50%
                      
                      
                      RL, address=17774400-17774411, vector=160, 4 units
                        RL0, 5242KW, attached to rtv53_rl.dsk, write enabled, RL02
                        RL1, 2621KW, not attached, write enabled, autosize
                        RL2, 2621KW, not attached, write enabled, autosize
                        RL3, 2621KW, not attached, write enabled, autosize
                      
                      boot rl0
                      
                      
                      RT-11FB  V05.03
                      
                      .TYPE V5USER.TXT
                      
                                                    RT-11 V5.3
                      
                             Installation of RT-11 Version 5.3 is complete and you are now
                          executing from the working volume    (provided you have used the
                          automatic installation procedure). DIGITAL recommends you verify
                          the correct  operation  of  your  system's  software  using  the
                          verification procedure.  To do this, enter the command:
                      
                                                   IND VERIFY
                      
                              Note that VERIFY should be performed  only after the distri-
                          bution media have been backed up.  This was accomplished as part
                          of automatic installation on  all  RL02,  RX02,  TK50, and  RX50
                          based systems,   including the  MicroPDP-11 and the Professional
                          300.  If you have not completed automatic installation, you must
                          perform a manual backup before using VERIFY.  Note also,  VERIFY
                          is NOT supported on RX01 diskettes,    DECtape I or II,   or the
                          Professional 325.
                      
                          DIGITAL also  recommends  you  read  the  file V5NOTE.TXT, which
                          contains information  formalized too late to be included  in the
                          Release Notes.  V5NOTE.TXT can be TYPED or PRINTED.
                      .
                      or
                      PDP-11 simulator V3.7-3 [32b data, 32b addresses, no Ethernet]
                      
                      Disabling XQ
                      CPU, 11/45, FPP, idle disabled, autoconfiguration on, 256KB
                      Throttle = 80%
                      
                      
                      RL, address=17774400-17774411, vector=160, 4 units
                        RL0, 5242KW, attached to unix_v7_rl.dsk, write enabled, RL02
                        RL1, 2621KW, not attached, write enabled, autosize
                        RL2, 2621KW, not attached, write enabled, autosize
                        RL3, 2621KW, not attached, write enabled, autosize
                      
                      boot rl0
                      boot
                      rl(0,0)rl2unix
                      
                      @boot
                      New Boot, known devices are hp ht rk rl rp tm vt
                      : rl(0,0)rl2unix
                      mem = 177856
                      # ls -la
                      total 644
                      drwxrwxrwx 8 root      240 Sep 22 05:47 .
                      drwxrwxrwx 8 root      240 Sep 22 05:47 ..
                      drwxrwxr-x 2 bin      2512 Sep 22 05:32 bin
                      -rwxr-xr-x 1 bin      8986 Jun  8  1979 boot
                      drwxrwxr-x 2 bin       160 Sep 22 05:47 dev
                      drwxrwxr-x 2 bin       336 Sep 22 05:33 etc
                      -rwxr-xr-x 1 sys     53302 Jun  8  1979 hphtunix
                      -rwxr-xr-x 1 sys     52850 Jun  8  1979 hptmunix
                      drwxrwxr-x 2 bin       320 Sep 22 05:33 lib
                      drwxrwxr-x 2 root       96 Sep 22 05:46 mdec
                      -rwxr-xr-x 1 root    50990 Jun  8  1979 rkunix
                      -rwxr-xr-x 1 root    51982 Jun  8  1979 rl2unix
                      -rwxr-xr-x 1 sys     51790 Jun  8  1979 rphtunix
                      -rwxr-xr-x 1 sys     51274 Jun  8  1979 rptmunix
                      drwxrwxr-x11 root      176 Sep 22 05:46 usr
                      #
                      # man ls
                      LS(1)               UNIX Programmer's Manual                LS(1)
                      NAME
                           ls  -  list contents of directory
                      ..........................................................................
                      Printed 9/22/88                                                 1
                      # ls -la bin
                      total 2593
                      drwxrwxr-x 2 bin      2512 Sep 22 05:32 .
                      drwxrwxrwx 9 root      256 Sep 22 05:50 ..
                      -rwxr-xr-x 1 bin       185 Jun  8  1979 1
                      -rwxr-xr-x 1 bin      7604 Jun  8  1979 ac
                      -rwxr-xr-x 1 bin     34652 Jun  8  1979 adb
                      -rwxr-xr-x 1 bin      9844 Jun  8  1979 ar
                      -rwxr-xr-x 1 bin      3610 Jun  8  1979 arcv
                      -rwxr-xr-x 1 bin      5604 Jun  8  1979 as
                      # tail -1 /etc/passwd
                      dmr::7:3::/usr/dmr:
                      It's really true Unix.
                      • 38. Re: Is ORACLE slower on Windows then on Linux ?
                        Billy~Verreynne
                        jgarry wrote:
                        It's because all the people without jobs can get their hands on is Linux.

                        Linux is popular because it's available to the great unwashed, that's the only reason . . .
                        Well, I may not agree with other things he says or his manner of saying them, but I don't agree that this is bigoted, and I believe to a large extent he is correct.
                        Of course it is bigoted... what does [+great unwashed+| http://en.wiktionary.org/wiki/the_great_unwashed] implies? A contemptuous term for the populace, particularly the working class.

                        An elitist and egotistical and discriminatory attitude against the so-called "+unwashed+" that do not have the same opportunities, same access to so-called "real" Unix, cannot afford to attend Harvard/Stanford/whatever, etc...

                        And I take great exception to a bigotry and discriminatory view like that.
                        • 39. Re: Is ORACLE slower on Windows then on Linux ?
                          Billy~Verreynne
                          jgarry wrote:

                          I'm not even going to address the memory management mess, except to note, in linux it has changed rapidly and still hasn't been gotten right for Oracle databases. Figuring that out is an exercise for the student. Have fun going up and down on the elevator in the swampiness, I mean, swapiness.
                          Very vague. Would like to see some hard technical fact and evidence to back that up that Linux memory management is inferior and not suited for databases. BTW, I run Oracle on Solaris, HP-UX and Linux in production.
                          The driver model sucks completely, you just don't know what is going to work when with what and for how long. Even freakin' winblows, which I'm strongly biased against, is better for that.
                          Again, provide technical fact. Linux does not have a monolithic driver stack and allows drives to be dynamically inserted and unloaded from the kernel. This allows me to troubleshoot driver and hardware problems without rebooting the kernel - as I can unload a complete driver stack, make config changes and reload the driver stack (or even a different version of that driver stack).

                          What happens if you get a driver failure on Windows? Or HP-UX? Can you rectify that without rebooting? Most cases not.

                          And you consider that a better approach? I don't.

                          Also remember that the Linux kernel developers are actual real-world Linux kernel users. They use it for some of the biggest computing clusters on this planet. They use it as web servers, app servers, database servers, network routers, firewalls, etc. Real world requirements find their way very quickly into the Linux kernel. Problematic features are equally fast removed (if it ever make it that far into the stable tree).

                          Linux is the most up to date and "applicable" kernel for running production business servers. Period.
                          • 40. Re: Is ORACLE slower on Windows then on Linux ?
                            706417
                            Billy, Billly, Billy... "The Great Unwashed" has many meanings. Trust a berk like you to selectively pick one meaning and then parade it about like a big pink poodle.


                            Regards - Don Lewis

                            P.S. Is your real surname Nomates?
                            • 41. Re: Is ORACLE slower on Windows then on Linux ?
                              Niall Litchfield
                              burleson wrote:
                              Ones we've installed Oracle on Windows 2003 Server and it was really slower like 200-300% .....
                              That's pretty good, I've seen a lot worse!

                              Windows is one of the most poorly designed operating systems ever invented, I don't recommend it for anybody . . .

                              I have justification here:

                              http://www.dba-oracle.com/t_bad_poor_performance_upgrade_migration_32_64_bit.htm
                              errm No you don't. You have an article about a migration from 32bit to 64bit on another platform.

                              >
                              Move to Linux, you won't regret it!
                              Niall Litchfield
                              http://www.orawin.info/
                              • 42. Re: Is ORACLE slower on Windows then on Linux ?
                                Niall Litchfield
                                Charles Hooper wrote:
                                Kris,

                                There are a lot of factors which may have contributed to the slower than expected performance on Windows 2003 that you noticed. Those factors include:
                                * Not collecting system (CPU) statistics on Oracle 9i or higher
                                * Not collecting data dictionary statistics on Oracle 9i or higher
                                * Not configuring large page support, if it is able to be used
                                * Not taking advantage of a battery backed cache to improve write performance
                                * Leaving the default db_file_multiblock_read_count set to 16 (I believe that is what the DBCA sets on 10g and above) rather than allowing Oracle to auto-tune the parameter. Limiting extent sizes to smaller than 1MB could also cause performance problems.
                                * Installing a virus scanner on the server, especially if it is permitted to scan program and data files used by Oracle.
                                * Using inappropriate parameters for memory allocations.
                                * If a database release upgrade was involved, not directly attacking the specific performance problem which is the source of the problem - could just be just a couple bad execution plans, possibly caused by the upgraded query optimizer.
                                * ... (Niall Litchfield, or someone else might be able to provide additional causes)
                                Probably the most common reasons are running excessive non-oracle workload on the same server (of which the virus scan is a prime example), but often people I see will compare a combined database and app server running windows against the same system but with the database remotely located on a unix server. The performance difference is then attributed to the o/s and not the architecture.

                                >
                                A couple things to keep in mind about Windows:
                                * Oracle on Windows uses direct, asynchronous I/O, which helps minimize concurrency problems. But, this also means that the operating system's file cache probably is of little use to Oracle on Windows.
                                conversely Linux, especially older releases of Oracle on Linux, tended to have really poor disk io in comparison to windows. This problem has mostly gone away, but for all its faults NTFS remains a really good FS for Oracle.

                                * Oracle on Windows uses a thread model, while Oracle on Linux (and other Unix like operating systems) uses a process model. The thread model is less negatively impacted by context switches than is the process model, which should help improve performance in some cases, such as repeatedly calling a PL/SQL function from a SQL statement.
                                The single biggest limitation these days however is in addressable memory on the 32bit platform, since all those threads share the same memory space (especially for PGA and session info) as the backgrounds. The simple answer to this is not to use PAE or /3gb or all the rest but to use a 64bit o/s (including windows).
                                * Properly configured, the same server running 64 bit Windows should be able to achieve roughly the same performance as the same server running 64 bit Linux (assuming direct, asynchronous I/O is enabled on Linux).
                                * ... (Niall Litchfield, or someone else might be able to provide additional things to keep in mind)
                                The tests that I and others have done (*in the 2.4 kernel timeframe*) showed that workload based tests (as opposed to Roby's operation based tests) gave windows about a 5-15% advantage when properly configured. It's high time these tests were re done though.

                                >
                                The general advice that you will likely receive is to use the operating system that you know best for Oracle.
                                That would be mine, technology doesn't break databases, people do.


                                Niall
                                • 43. Re: Is ORACLE slower on Windows then on Linux ?
                                  21628
                                  >
                                  Man-oh-man . . .

                                  Sorry, I absolutely detest spelling Nazis . . .

                                  I spent my college time learning computer science, not memorizing spelling . . .

                                  Personally, I think that people who fixate on trivial non-issues (spelling, grammar, typos) are rude, not paying attention to the arguments. . . .
                                  >

                                  Huh. I would normally expect that someone who emphasizes dress codes ("A professional consultant who doesn’t take the time to maintain a professional appearance presents the image of not being able to perform adequately on the job" - Burleson's web site) would recognize that sloppy spelling could well indicate a lack of professionalism. My impression, however, from reading through lots of Burleson's crap over the years, is that he is not someone I would describe as professional anyway. A hand-waving, self-promoting charlatan, maybe.
                                  • 44. Re: Is ORACLE slower on Windows then on Linux ?
                                    Billy~Verreynne
                                    Niall Litchfield wrote:
                                    burleson wrote:
                                    Windows is one of the most poorly designed operating systems ever invented, I don't recommend it for anybody . . .

                                    I have justification here:

                                    http://www.dba-oracle.com/t_bad_poor_performance_upgrade_migration_32_64_bit.htm
                                    errm No you don't. You have an article about a migration from 32bit to 64bit on another platform.
                                    Why even bother Niall..?

                                    It is very obvious that Burleson is clueless about operating system design and that any advice that he offers in this regard is based on hearsay, personal bias and total ignorance - and such meaningless and worthless advice is best ignored.